Molecular Gastronomy: Exploring the Science of Flavor (19 page)

selected, behind pasta, rice, and fried potatoes. The most popular dishes were

French fries, small sausages of the kind known in France and elsewhere in

Europe as chipolatas, quiches, pasta, breaded fish, rice, mashed potatoes, ham,

and beefsteak. No surprise, any of this—only a quantified confirmation of what

has long been observed.

On the other hand, the choice of meats and vegetables did hold surprises

for the researchers. The children made little distinction between them, choos-

ing roast pork, turkey, leg of lamb, and organ meats almost indiscriminately.

But when it came to vegetables, which were chosen less often than meats,

preferences were quite clear, depending on the type of vegetable and the type

of culinary preparation. Spinach, which is so widely thought to be disliked by

children, was selected more often and consumed in greater quantities than any

other vegetable—as long as it was napped with a white sauce. Endives, cab-

bage (raw or cooked), tomatoes, and green beans appealed to few of the young

diners, however. It appears that foods with a hard and fibrous texture (which

makes them hard to chew) are more likely to be refused, as are those with a

pronounced bitter flavor.

How are we to account for all this? It occurred to the sensory biochemists

in Dijon to try to relate the selection of foods to their nutritional content. What

At Table in the Nursery
| 119

they discovered was that the higher the caloric value of a food, the more often it

is chosen (with the exception of cheeses). Is this proof that innate dispositions

are still strong in young children? We know that infants who are fed salt, sweet,

sour, or bitter liquids react visibly with disgust in the case of the last two flavors

but show pleasure when they encounter a sweet taste. In this they resemble

fruit-eating monkeys, who associate sweetness with the presence of sugars

(whose molecules have a high energy value) and bitter with vegetables that

contain toxic alkaloids. Gradually children come to learn, through conditioning

and culture, to diversify their diet. Conditioning causes children to associate

satiety, for example, with the consumption of high-calorie nonsweet dishes

(fatty foods in particular); culture accounts for the fact that some children, even

when quite young, learn to relish dishes with a powerful taste, contrary to what

purely biological explanations would predict.

All Equal, All Di‡erent

None of the factors analyzed so far explains the highly variable choice of

foods with a strong taste—neither sex, nor maternal feeding, nor the child’s

birth order, nor his or her body mass index (a measure of fatty mass obtained

by dividing weight by the square of the child’s height).

Because the dietary preferences of young children have been little studied

until now, it is not surprising that these first studies have left many observa-

tions unexplained. Moreover, it will be a long time before we are able to corre-

late the choices of the children studied with their preferences as adults. But pa-

tience is the price we must pay if we are to identify the mechanisms of dietary

learning in childhood and to clarify the role of early exposure to different foods

in the formation of later preferences—long thought to be important but so far

undemonstrated. The food industry takes an obvious interest in the results of

these studies, for it will be able to offer products of longer-lasting appeal only if

the dynamics of the dietary preferences of young children are understood.

120 | t he physiology of f l a vor

33

Food Allergies

Predicting and preventing the risks of allergic reaction to transgenic

foods.

g a s t r o n o m y i s n o t a w o r l d o n l y of pleasurable aromas and tastes.

More than one-quarter of the population in seven countries of the European

Union claims to suffer from food allergies or intolerances. Although clinical

tests indicate that the actual incidence of such allergies is much lower than

commonly believed (about 3.5% of the population), these reactions pose a

major public health problem, all the more because the seriousness of the at-

tacks reported is on the rise. In the last ten years the number of cases of ana-

phylactic shock caused by a food allergy has quintupled, and many of these

attacks are fatal, notably ones triggered by the ingestion of products derived

from peanuts.

Jean-Michel Wal and his colleagues at the Institut National de la Recherche

Agronomique-Commissariat de l’Énergie Atomique (inra-cea) Laboratoire

d’Immuno-Allergie Alimentaire in Saclay have shown how a protein found in

cow’s milk triggers allergies despite its resemblance to a human protein that

is readily tolerated. They are also conducting animal tests of a method of gene

immunization for preventing allergies.

The reasons for the recent growth of food allergies are not completely un-

derstood. The increase in the number of offending foods seems to be linked

to the development of respiratory allergies to plant pollens; the antibodies pro-

duced by the human organism against a pollen antigen sometimes react also

against molecules from a very different source. These cross-reactions are a

| 121

product of shared epitopes, molecular fragments that are the target of the im-

mune system. Thus the multiplication of allergies to exotic fruits (avocados,

kiwi fruits, bananas, and so on) has been found to be associated with the de-

velopment of allergies to latex, themselves on the rise as well.

To these traditional foods have been added new ones derived from geneti-

cally modified organisms, which contain proteins normally absent in tradition-

al foods. The growing commercial availability of such products has aroused

concerns that they may be allergenic, but because few cases of illness have

been reported until now it is not clear how to assess the risk posed by the

introduction of new foods. Several approaches are being investigated: tests

involving serums from allergic patients (whose antibodies bind to allergenic

molecules), cutaneous tests (in which molecules thought to be allergenic are

deposited on the skin, and signs of adverse reaction are monitored), animal

studies (in which candidate molecules are administered to sensitized animals),

and theoretical models of the incidence of allergic reactions based on the pro-

tein sequence in amino acids.

Developing such models will take time, but the preliminary step of com-

piling sets of well-characterized data for proteins whose allergenic character

is known is now being carried out. Wal and his colleagues have already used

these databases to compare the immune reactivity of human and bovine beta-

casein.

Allerês to Milk

Caseins constitute 80% of the proteins found in milk. There are four types:

alpha S1, alpha S2, beta, and kappa. The beta-casein in cow’s milk is an im-

portant allergen: 92% of the serums obtained from people allergic to caseins

contain a type of antibody responsible for allergic reactions, known as immu-

noglobulin E, that is specifically directed against beta-casein.

In 1997 the immunologists at Saclay demonstrated the presence of immu-

noglobulin E molecules specific to a bovine whey protein in the blood of aller-

gic people and observed that these molecules also react with the corresponding

human protein. Does this same kind of cross-reaction exist for beta-casein?

Wal and his colleagues tested the serum of twenty people allergic to the beta-ca-

sein of cow’s milk, carefully measuring the concentration of immunoglobulin

122 | t he physiology of f l a vor

E specific to bovine beta-casein and then verifying that these molecules also

reacted with human beta-casein.

Human and bovine beta-caseins share 50% of their amino acids, hence the

interest in studying their common domains. One such domain corresponds

to an amino acid sequence that assumes a helical form in both human and

bovine proteins. Another common domain contains the principal phosphory-

lation sites of beta-casein. Once the amino acids that make up this protein

have been assembled, it is modified in particular by the addition of phosphate

groups, a process known as phosphorylation. Such posttranslational chemical

modifications influence the properties of the molecules by changing their elec-

trical charge and altering their conformation. The inra-cea immunologists

were able to show that the phosphorylation of this second domain affects the

allergenic character of beta-casein.

Preventive Measures

Preliminary studies of new foods derived from genetically modified organ-

isms have not yet shown any of them to cause allergies. However, research is

being conducted on possible preventive measures. In the case of beta-casein,

milk substitution has proved to be ineffective because the caseins of different

kinds of milk are very similar. The Saclay team therefore turned its attention

to testing gene immunization in mice. This involves injecting the animals with

bacterial dna containing additional genetic material that codes for a particular

protein. These sequences activate a nonallergic reaction of the immune system

that switches off the allergic reaction. In the case of allergies to bovine beta-

lactoglobulin (another protein found in cow’s milk), the method was found

to substantially and persistently reduce the production of immunoglobulin E

specific to this protein.

Food Allergies
| 123

34

Public Health Alerts

Food warnings attest to the progress being made by microbiolo³.

c a s e s o f l i s t e r i o s i s a r e d r a m a t i c, sometimes fatal, and all the more

shocking because measures to protect public health have never been more ef-

fective. Nonetheless this relative security cannot be counted on to prevent the

occurrence of illness, even death, among people in the leading risk groups:

pregnant women, the elderly, young children, and those with immunodefi-

ciencies. Because health alerts are issued as a consequence of retrospective

epidemiological inquiries that compare strains isolated from the victims with

products suspected of contamination, they constitute a response to crisis rath-

er than a means of preventing it. One of the profound challenges facing basic

research in microbiology is to develop a better understanding of
Listeria
and

other food microorganisms.

The most urgent task is to identify pathogenic strains. Only when these are

known can the populations at risk be determined and a range of commercially

available foods tested. At the Institut National de la Recherche Agronomique

Laboratoire de Pathologie Infectieuse et Immunologie at Nouzilly, Patrick

Pardon and his colleagues are studying the comparative virulence of
Listeria
.

The food-related strains of this bacterium that have been isolated have vari-

able pathogenicities. How can dangerous strains be distinguished from benign

ones? In recent years microbiologists have distinguished two types:
Listeria

monocytogenes,
considered always to be pathogenic, and a harmless variety

called
Listeria innocua
. But even this classification, which has made it possible

124 |

to avoid issuing unnecessary alerts, is inadequate. The problem is that present-

day techniques do not allow the virulence of different strains to be character-

ized with precision.

Rapid Testing for Necessary Alerts

The virulence of different strains has long been studied in laboratory ani-

mals (chiefly mice) inoculated with microorganisms. This method is both time

consuming and costly, and it should be abandoned once alternative procedures

become available.

The Nouzilly microbiologists developed an in vitro test that uses human

intestinal cells, which offer a point of entry to the organism. To test the patho-

genicity of a particular strain of
Listeria,
one places a sample in suspension on

a culture of these cells. One then looks to see whether ruptures appear in the

cell layer in response to the introduction of the bacteria.

By means of this test the biologists were able to discover that the
Listeria

monocytogenes
group is not homogeneous: Certain strains were virulent to only

a small degree or not at all (the results were confirmed on laboratory animals),

which led to further analysis of the genome and the proteins of these bacteria.

A three-year national program is under way in France in which several teams

of researchers are reviewing a database of about 400 strains of
Listeria mono-

cytogenes
. The genes and proteins of strains that pose only a slight pathogenic

risk, or none at all, will then be compared with those of virulent strains in the

hope of further refining the detection of pathogenic strains.